Conversational commerce and the power of the bot

Bots are not new, but they are bursting into the marketing buzzword library with renewed vigour. By its simplest definition a bot is “a computer program designed to simulate conversations with human users, especially over the internet”, the execution of a ‘bot’ is still blurry and encompasses a number of different interpretations.

There are simple stand-alone conversational interfaces like the Quartz App or complex digital assistants like Siri and Alexa; natural language processing (NLP) tools such as wit.ai, cognitive technologies such as IBM’s Watson and bot developer frameworks like Facebook’s Messenger Bot API. With so many big players willing to dive into this space, we are seeing more and more brands willing to dip their toes into the pool in order provide a service to consumers that is personal, contextual and cost efficient.

Applications and websites serving blanket content in traditional ways have proven to be unfruitful, return rates are poor and ROI is questionable. Contextual relevance is of the utmost importance, utility needs to be prioritized. Automating recurring and repetitious customer service tasks can reduce costs and improve responsiveness, or perhaps even, be proactive without additional expense.

Using existing platforms like Facebook Messenger enables brands to piggyback their way into consumer’s phones without a stand alone application, using this platform their target market are already in and pick up the conversation. Developers can utilise existing development frameworks, streamlining the path to MVP. Behind a conversational interface, there are two potential development approaches- integration with an Artificial Intelligence (AI) system or a predefined list of rules and programmed conversation that do not let the users diverge from an option of paths.

With AI being in infancy and implementation at a hefty price tag, the ‘decision tree’ method whilst by no means the only way, is the quickest and most cost effective way to get into market. By extension, the conversation can only be as complex as it is programmed to be. The current capacity of the tech means that we are limited in the effectiveness of free form text — the complexity of naturalised language is a difficult beast to master. Succinct interactions (buttons/suggested replies) comprised of a series of straightforward processes are the most achievable and unassuming methods of bot communication. The outcomeis the ability to close the loop with the customer along a process, efficiently and in a delightful way. The ROI is in cost reduction, efficiency and improved customer satisfaction.

This may sound like cutting out the heart of a creative execution, but there is still a lot to be learnt from even the most simple bots. Our ability to effectively A/B test creative execution, tone of voice and extract information of contextual relevance to the user provides us with deeper understanding of the customer journey and emotional lifecycle. From here effective and contextual content and conversation can be successfully mapped, programmed and learned from.

As we are all aware, game changers in technology come hard and fast- so whilst at this moment in time the ROI on complex AI bots may prove a step too far, we need to be mindful that we are on the cusp of normalising a new way to interface with digital systems- even with the most basic of executions.

We can potentially bridge the gap between digital and physical spaces by exploiting something core to our nature; our ability to connect through language and our inclination for seeing human features in the non-human scenarios: anthropomorphisation.

Whilst this is no doubt a manipulation, what makes bots (both simple and complex) effective is our ability to develop a psychological bridge with consumers. Breaking down mechanical interaction, creating connection and seemingly empathy and strengthening the brand/consumer bond.

The psychological effectiveness of anthropomorphizing digital systems is not a new phenomenon. Arguably the first foray into this area of research proved our human flaw in meaningfully connecting with anything that shows the faintest human sentiment. In 1963, Joseph Weisenbaum at MIT created non-directive Rogerian psychoanalysis tool called ELIZA. The simple algorithm would detect key words in the users sentences and echo them back in an attempt to delve the conversation further.The experiment was cancelled when the project divisor realized that multiple staff members sneaking into the lab to talk to ELIZA after hours for time alone with the machine. As Weisenbaum noted “I had not realized … that extremely short exposures to a relatively simple computer program could induce powerful delusional thinking in quite normal people.” Weizenbaum, Joseph (1976). Computer power and human reason: from judgment to calculation.

Attribution of human like features to computer based systems and our willingness to connect to them are arguably, a slippery slope. But these are realities that we will continue to face as we continue to the path of AI, automation and conversational interfaces. As we welcome systems like Siri, Alexa and Google Assistant into our homes and lives, our literacy of these systems will soon become our expectation when dealing with brands.

Why does this matter? It is important that we begin to familiarise ourselves and interact with the interfaces that are just over the horizon — and bots are on the front line of this transition. Yes, we will continue having teething issues as AI comes into maturity, but getting despondent over setbacks is not effective. Seeing this as a ‘fad’ is shortsighted. Whilst the platforms, frameworks and approach may change, conversational and cognitive commerce is definitely here to stay.

Kara Bombell, Mobile and Production Director

Show your support

Clapping shows how much you appreciated Isobar Australia’s story.